forked from wjn1996/DaseRecSys
-
Notifications
You must be signed in to change notification settings - Fork 0
/
model.py
168 lines (140 loc) · 5.74 KB
/
model.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
import pandas as pd
import numpy as np
import json
import csv
import os
# 评分预测 1-5
class MatrixDecomForRecSys(object):
def __init__(
self,
lr,
batch_size,
reg_p,
reg_q,
hidden_size=10,
epoch=10,
columns=["uid", "iid", "rating"],
metric=None,
):
self.lr = lr # 学习率
self.batch_size = batch_size
self.reg_p = reg_p # P矩阵正则系数
self.reg_q = reg_q # Q矩阵正则系数
self.hidden_size = hidden_size # 隐向量维度
self.epoch = epoch # 最大迭代次数
self.columns = columns
self.metric = metric
def load_dataset(self, train_data, dev_data):
self.train_data = pd.DataFrame(train_data)
self.dev_data = pd.DataFrame(dev_data)
self.users_ratings = train_data.groupby(self.columns[0]).agg([list])[[self.columns[1], self.columns[2]]]
self.items_ratings = train_data.groupby(self.columns[1]).agg([list])[[self.columns[0], self.columns[2]]]
self.globalMean = self.train_data[self.columns[2]].mean()
def _init_matrix(self):
'''
*********************************
用户矩阵P和物品矩阵Q的初始化也对算法优化有一定帮助,更好的初始化相当于先验信息。
加分项:
- 思考初始化的一些方法,正态分布等等;
- 其他初始化方法?
*********************************
'''
# User-LF
P = dict(zip(
self.users_ratings.index,
np.random.rand(len(self.users_ratings), self.hidden_size).astype(np.float32)
))
# Item-LF
Q = dict(zip(
self.items_ratings.index,
np.random.rand(len(self.items_ratings), self.hidden_size).astype(np.float32)
))
return P, Q
def train(self, optimizer_type: str):
'''
训练模型
:param dataset: uid, iid, rating
:return:
'''
P, Q = self._init_matrix() # 初始化user、item矩阵
best_metric_result = None
best_P, best_Q = P, Q
for i in range(self.epoch):
print("Epoch: %d"%i)
# 当前epoch,执行优化算法:
if optimizer_type == "SGD": # 随机梯度下降
P, Q = self.sgd(P, Q)
elif optimizer_type == "BGD": # 批量梯度下降
P, Q = self.bgd(P ,Q, batch_size=self.batch_size)
else:
raise NotImplementedError("Please choose one of SGD and BGD.")
# 当前epoch优化后,在验证集上验证,并保存目前最好的P和Q
metric_result = self.eval(P, Q)
# 如果当前的RMSE更低,则保存
print("Current dev metric result: {}".format(metric_result))
if best_metric_result is None or metric_result <= best_metric_result:
best_metric_result = metric_result
best_P, best_Q = P, Q
print("Best dev metric result: {}".format(best_metric_result))
# 最后保存最好的P和Q
np.savez("best_pq.npz", P=best_P, Q=best_Q)
def sgd(self, P, Q):
'''
*********************************
基本分:请实现【批量梯度下降】优化
加分项:进一步优化如下
- 考虑偏置项
- 考虑正则化
- 考虑协同过滤
*********************************
'''
return P, Q
def bgd(self, P, Q, batch_size: int=8):
'''
*********************************
基本分:请实现【批量梯度下降】优化
加分项:进一步优化如下
- 考虑偏置项
- 考虑正则化
- 考虑协同过滤
*********************************
'''
return P, Q
def predict_user_item_rating(self, uid, iid, P, Q):
# 如果uid或iid不在,我们使用全剧平均分作为预测结果返回
if uid not in self.users_ratings.index or iid not in self.items_ratings.index:
return self.globalMean
p_u = P[uid]
q_i = Q[iid]
return np.dot(p_u, q_i)
def eval(self, P, Q):
# 根据当前的P和Q,在dev上进行验证,挑选最好的P和Q向量
dev_loss = 0.
prediction, ground_truth = list(), list()
for uid, iid, real_rating in self.dev_data.itertuples(index=False):
prediction_rating = self.predict_user_item_rating(uid, iid, P, Q)
# dev_loss += abs(prediction_rating - real_rating)
prediction.append(prediction_rating)
ground_truth.append(real_rating)
metric_result = self.metric(ground_truth, prediction)
return metric_result
def test(self, test_data):
'''预测测试集榜单数据'''
# 预测结果可以提交至:https://www.kaggle.com/competitions/dase-recsys/overview
test_data = pd.DataFrame(test_data)
# 加载训练好的P和Q
best_pq = np.load("best_pq.npz", allow_pickle=True)
P, Q = best_pq["P"][()], best_pq["Q"][()]
save_results = list()
for uid, iid in test_data.itertuples(index=False):
pred_rating = self.predict_user_item_rating(uid, iid, P, Q)
save_results.append(pred_rating)
log_path = "submit_results.csv"
if os.path.exists(log_path):
os.remove(log_path)
file = open(log_path, 'a+', encoding='utf-8', newline='')
csv_writer = csv.writer(file)
csv_writer.writerow([f'ID', 'rating'])
for ei, rating in enumerate(save_results):
csv_writer.writerow([ei, rating])
file.close()